Sometimes the hardest part of a radical business transformation is getting people to let go of what the company used to do.
McGraw-Hill was founded in 1888 by James McGraw and John Hill, and 100 years later it was the largest educational publisher in the United States. But as with many traditional publishers, that success eroded in the digital age. By 2013, when it was clear the business of publishing printed textbooks had no future, the McGraw-Hill Companies, now S&P Global, sold McGraw-Hill Education to Apollo Global Management for $2.4 billion.
What happened next was a notable change of direction. McGraw-Hill Education began a critical pivot to adopt technology and software to reimagine education.
A year later, David Levin became CEO, with a charge to lead the transition and define the company around harnessing learning science with a new vision of unlocking the full potential of each learner. This was based on its new adaptive technology built around established theories such as mastery learning. In the four years since it became independent, the company has spent more than $700 million on its organic digital investments and technology acquisitions.
We interviewed Levin from his New York City office.
During the three years that you've been CEO, you've been leading a transition to a completely different kind of business. What challenges did you face, and how has talent development contributed to meeting them?
We have to deal with a changing business model, new customer needs and demands, and new skill sets and capabilities—and with the perception of being a legacy business. So, if you come into something with the context of radical change, your job is to establish a credible path to the future; deploy resources to make it happen; and use that to give hope and substance to the promise, both internally and externally.
That has involved reflecting on strategy, but also embedding strategy in the core leadership group. A huge part of that depends on talent development because we need people who can understand and grow the new business and develop and amplify a very different message to the market.
We've also put real resources to work, and since the spin-off, we've invested in creating a high-end engineering capability in-house. Our digital product group now stands at just over 500 really talented people who are shaping the future.
Given that kind of energy, effort, and resources, the need to motivate and support the talent that you're recruiting, developing, and retaining is profound. That means having strategies to support their professional engagement and to shape the culture of a company that will be doing very different things with a changing workforce.
How did you get those new core competencies into the company?
I was blessed in a particular way. My predecessor had just recruited a brilliant individual to head the digital platform group. I spent a lot of time talking with him through my recruitment, as I was skeptical of joining as CEO because I thought the extent of the challenge was so enormous. We really had a mind meld around values and software strategy, and those conversations convinced me that we had a linchpin around whom we could build a core competency in engineering.
He was—and is—a great leader who could create the right framework of values and processes, and who also had the ability to execute a software strategy. I knew we could deliver by providing the resources and the strategic alignment in the other parts of the business around this core.
What did you see missing from the culture? What needed to change?
We were a fabulous old company, and for about the first 125 years, everyone understood our business to be publishing. In the eyes of the public, we were a textbook publisher. Textbook publishing is built on two pillars: content and a transactional sales model. That's the model that has powered publishing from Gutenberg until about three or four years ago. But with software, there is a need for change. So in sales, we're moving toward skill sets that are about long-term relationships and a culture of customer success as opposed to transactional selling.
We're moving to a business which is defined by our ability to influence student outcomes directly. That means we need to have the ability to actually support and help drive better results for students through the creation of data and analytics on their usage of content and engagement with it.
And while content remains incredibly important, we also have to acknowledge that ours is not the only content that people want to use, so we've had to open our software to third-party content and allow people to add to and supplement our programs. Five years ago, that would have been heresy.
The new McGraw-Hill Education describes itself as a learning sciences company. What does that mean?
We look at learning science as an evolving practice which harnesses the best of psychology and cognitive science to understand how people learn, and applies this in software to create products that enhance student learning and support teachers in their critical roles of helping students.
In our software, we are applying a variety of well-understood but still evolving reflections on how people learn, with the intent of using them to make learning work better.
Back in the 1890s, a German psychologist named Hermann Ebbinghaus used himself as an experiment to find the best process for remembering randomized strings of three-letter words. Some of these strings had up to 1,000 or more words. He determined how many repetitions, with what frequency, it took to move something into short-term memory and from there to long-term memory. This eventually became an established part of psychology known as effortful retrieval and spaced practice.
In the 1980s, the American educational psychologist Benjamin Bloom helped develop the theory of mastery learning. In very simple terms, mastery learning means you don't move on to new material until you've really learned the current material.
Bloom found that students at the bottom of the grading curve in classes using conventional teaching methods got higher grades in classes using mastery learning techniques. And their grades improved even more when they were given individual tutors.
In our software, we've created a structure to support mastery learning and to provide data about each student's level of mastery to indicate when and where they need tutorials. This layer of data and analytics allows a single instructor to know who's got what problem and to focus time exclusively on that.
So, when we talk about learning science, it's the combination of improving learning with mastery and competence-based learning techniques, and at the same time, embedding in our software some analytics that help people learn as fast and efficiently as possible. This also allows teachers or faculty to spot where they can best interact with students to make a difference. It's a profound shift for our company. Now our goal is to support great outcomes; not only great content, but great outcomes.
This is a game shift for learners too. When they have mastery, they're able to execute and perform at their best. And it's a game changer for talent development. For the first time we can look at the performance of an individual student or a cohort of students and say, "Here is the deficiency and here is what you need to teach." It's so much more efficient than teaching to one part of the curve and losing everyone else, or covering material without knowing if it's what learners actually need.
The beauty of this system of learning is that it allows the creation of time and space to focus on what's most important to know at every level of an organization.
Where do you think learning science is headed? What might be the next breakthrough?
We've now got some well-established understanding of how people learn better. Teachers and trainers are going to have access to more information about the learning state of their students and learners, and the real question will be how they make use of that most effectively.
I think the breakthrough that will happen over the next five years is around instruction and learning supported by data. How do we support changes in instruction—in professional development, in universities, in schools, and in the workplace—to make full use of all the insights that come out of data and learning analytics? That's massive. It can empower teachers to be the best they can be. It will help learners achieve more, faster.
How do you see the role of the trainer evolving in this environment?
The role changes from being the sage on a stage to being a coach, facilitator, and supporter who's actively intervening in the learning path of the students and engaging them in learning what they don't know. Metacognitive theory, which states that learners learn best when they know what they don't know, is one of the theories underlying our educational technology.
I see trainers becoming more efficient and effective and more focused in their use of time. This will have huge impacts as they deal with the ever-larger number of learners demanded by the knowledge economy.
What about changes to the role of the instructional designer?
Our framework says that instruction has to be intuitive, easy to use, engaging, efficient, and effective.
The nature of content has to shift too. Think of those ghastly compliance videos that we've all had to watch. Here's a typical scenario: "Do you bribe or do you not bribe? Click A or B."
What we really need is instruction that can handle complexity and drive mastery, and that can, for example, teach pharmaceutical salespeople to understand and convey the interaction between product A and product B.
What we look for in our design is to increase conscious competence. We want the system to support learners' confidence in their knowledge so that they knock the ball out of the park because they know exactly what they're doing.
We need to help build the confidence of a learner who's consistently right but not confident about it. And, in parallel, we need to address unconscious incompetence, which is the person who's really confident but consistently wrong and focus them on their learning gaps. Our systems look at confidence as distinct from accuracy, so they can spot both confidence and competence and give the appropriate feedback around each of them.
What about the need to interpret the data that systems such as yours provide? Will that job fall to the trainer or to someone else?
It's going to depend on the nature of the problem being solved, the organization that's deploying the software, and the criticality of mastery. For example, we looked at the partner group of a major accounting firm and could show very clearly the different levels of mastery of the tax code—a very complex subject—across the group.
In that situation, people at many levels would be interested in using the data: the professionals who would act as coaches, the trainers helping individuals learn according to their needs, the chief human resource officer, and even the CEO who would want the partners to be sufficiently skilled in the tax code as a matter of strategy. But for a frontline role, it could be more basic and dependent just on the trainer.
What about within McGraw-Hill Education itself? How is learning science being applied there?
We now use our technologies in our induction process. We're a complicated organization with 5,000-odd people spread across many countries. Our culture is evolving. What we sell is changing. So, it's a beautiful environment to use as a test bed for adaptive learning. We use it to help newcomers to the organization become effective really quickly.
Then we said, "Look, if we think this is so powerful for kids and students, why don't we make our products freely available for our own people and their children?" We've done that and usage rates are extraordinarily high. They know it works.
I've used it myself with my youngest son. When we moved to the States three years ago, he had to switch from the British school system to the American system. We used our learning software for math together. I did the high school math course alongside him as a competition, and it was a revelation: high school math is tough, I've got to say. We stopped the contest in 10th grade, but I continue to study polynomials where he now has a proficiency that I don't.
From a learning sciences perspective, do you see distinctions between knowledge mastery and application?
Yes, and there's a hierarchy. Some subjects require you to know a basic structure, but how to apply it is a separate subject.
A good example is learning a language. You have to know vocabulary and grammar with a high level of confidence in order to go forward. But knowledge of vocabulary and grammar is not sufficient to allow you to have a free-flowing conversation. This is where mastery learning becomes a bridge by reinforcing your confidence that you are competent, which makes you willing to try having a conversation.
In the workplace, learning science can support people in becoming more confident in their competence and thus more effective in their jobs.